Behav Ecol Sociobiol (2007) 61:689–701
DOI 10.1007/s00265-006-0299-5
ORIGINAL ARTICLE
Environmental and genetic influences on mating strategies
along a replicated food availability gradient in guppies
(Poecilia reticulata)
Gita R. Kolluru & Gregory F. Grether & Heidy Contreras
Received: 16 January 2006 / Revised: 23 October 2006 / Accepted: 28 October 2006 / Published online: 13 December 2006
# Springer-Verlag 2006
Abstract Food availability is expected to influence the
relative cost of different mating tactics, but little attention
has been paid to this potential source of adaptive
geographic variation in behavior. Associations between
the frequency of different mating tactics and resource
availability could arise because tactic use responds directly
to food intake (phenotypic plasticity), because populations
exposed to different average levels of food availability have
diverged genetically in tactic use, or both. Different
populations of guppies (Poecilia reticulata) in Trinidad
experience different average levels of food availability. We
combined field observations with laboratory “common
garden” and diet experiments to examine how this
environmental gradient has influenced the evolution of
male mating tactics. Three independent components of
variation in male behavior were found in the field:
courtship versus foraging, dominance interactions, and
interference competition versus searching for mates. Compared with low-food-availability sites, males at high-foodavailability sites devoted more effort to interference
competition. This difference disappeared in the common
garden experiment, which suggests that it was caused by
phenotypic plasticity and not genetic divergence. In the diet
Communicated by C. St. Mary
G. R. Kolluru (*) : G. F. Grether
Department of Ecology and Evolutionary Biology,
University of California, Los Angeles,
621 Charles E. Young Drive South,
Los Angeles, CA 90095-1606, USA
e-mail: [email protected]
Present address:
H. Contreras
Department of Ecology and Evolutionary Biology,
University of California, Irvine,
Irvine, CA 92697-2525, USA
experiment, interference competition was more frequent
and intense among males raised on the greater of two food
levels, but this was only true for fish descended from sites
with low food availability. Thus, the association between
interference competition and food availability in the field
can be attributed to a genetically variable norm of reaction.
Genetically variable norms of reaction with respect to food
intake were found for the other two behavioral components
as well and are discussed in relation to the patterns
observed in the field. Our results indicate that food
availability gradients are an important, albeit complex,
source of geographic variation in male mating strategies.
Keywords Mating tactic . Resource availability .
Geographic variation . Intrasexual competition .
Dominance . Phenotypic plasticity .
Genotype by environment interaction
Introduction
Intraspecific variation in mating strategies is well documented in a variety of taxa (Gross 1996; Foster and Endler
1999; Brockmann 2001; Shuster and Wade 2003). However,
only a few studies have linked geographic variation in
alternative mating behaviors to specific environmental
factors other than predation intensity (e.g., Carroll 1993;
Carroll and Corneli 1995). Food intake is particularly
important in influencing mating behavior (Blanckenhorn et
al. 1995; Belovsky et al. 1996; Plaistow and Siva-Jothy
1996; Moczek and Emlen 1999), and food availability
gradients offer excellent opportunities to study geographic
variation in mating tactics (Carroll and Corneli 1999). A key
step to understanding variation along environmental gradients is determining whether the differences among
690
populations represent phenotypic plasticity or local genetic
adaptation (Carroll and Corneli 1999; Foster and Endler
1999; Weitere et al. 2004). In this paper, we use a
combination of field observations and common-environment
lab experiments to tease apart these alternatives in guppies
(Poecilia reticulata).
Guppies are a classic system for studies of geographic
variation in behavior (Houde and Endler 1990; Endler
1995; Houde 1997). Grether et al. (1999, 2001) documented a replicated food availability gradient among lowpredation guppy streams in Trinidad. Low-predation
streams contain no piscivorous fish except Rivulus hartii,
which only occasionally eats guppies and preys mainly on
juveniles (Endler 1978, 1995). The main source of food for
guppies in these streams is attached unicellular algae
(Dussault and Kramer 1981), the abundance of which is
largely a function of forest canopy cover. Sites that receive
more light have larger standing crops of algae, but not
correspondingly higher densities of guppies, than sites
within the same drainage that receive less light (Grether et
al. 2001). In the high-light, high-food-availability sites,
female and juvenile guppies grow faster and males mature
at larger asymptotic sizes than their counterparts in the lowfood-availability sites (Grether et al. 2001).
The mating tactics of male guppies, in increasing order
of presumed energetic investment, include sneaking copulations without courtship display, displaying to females
before copulation (courtship), and aggressively inhibiting
rival males (Rodd and Sokolowski 1995; Houde 1997;
Jirotkul 2000; Kelly and Godin 2001). Male–male aggression occurs in two basic contexts, which, for brevity, we
call “competition” and “dominance” (Kolluru and Grether
2005). “Competition” refers to interference between males
simultaneously attempting to court the same female,
whereas “dominance” refers to aggressive interactions
between males out of the immediate proximity of females.
The latter may serve to establish or maintain dominance
relationships with respect to priority of access to receptive
females.
In this paper, we present results from an extensive
survey of behavioral variation across ten guppy sites in the
Northern Range of Trinidad. Sites were chosen to represent
the available extremes in forest canopy cover, and hence
food availability, among low-predation streams, and were
classified as either “high” or “low” with respect to food
availability, using previously established criteria (see
Materials and methods). Detailed behavioral (focal) observations were carried out on individual males at each site
and recorded along with data on local conditions (sex ratio,
fish density, and light level). Offspring of wild-caught
females were raised under “common garden” conditions in
the laboratory and observed in mixed-sex tanks under
standardized conditions after reaching sexual maturity. Four
Behav Ecol Sociobiol (2007) 61:689–701
of the ten populations were included in an experiment on
the effects of lifetime food intake on male mating tactics.
Results from the latter experiment have been published
(Kolluru and Grether 2005) but are reanalyzed here for
comparison to the field observations.
We predicted that males in high-food-availability sites
would engage in energetically-demanding (aggressive)
mating tactics more frequently than males in low-foodavailability sites. If the field differences between low- and
high-food-availability sites were due to phenotypic plasticity, then any phenotypic differences between low- and highfood-availability sites in the field should largely disappear
in the lab. Alternatively, the differences may persist in the
lab-reared males, suggesting that the field differences are
due to genetic divergence between low- and high-foodavailability sites. A third possibility is that the extent to
which behavior is flexible in response to food intake varies
among populations (genotype by environment interaction;
Thompson 1999). Environmental factors other than food
availability may also influence mating strategies in the
wild. Light levels, density, and sex ratio have been shown
to influence guppy mating tactics (Reynolds et al. 1993;
Houde 1997; Jirotkul 1999a,b, 2000; Gamble et al. 2003).
Our main objective in this paper is to examine each of these
possible influences on geographic variation in the mating
tactics of guppies and evaluate their relative importance.
Materials and methods
Study sites
Study sites meeting the following criteria were chosen
during a survey of stream drainages conducted in 1996 and
2000: (1) intact primary or old secondary growth rainforest;
(2) relatively homogeneous forest canopy cover within
sites; (3) separated from each other by multiple barriers to
guppy dispersal, including two or more waterfalls; and (4)
no predatory fish except R. hartii. We chose one stream
with relatively low canopy openness and another with
relatively high canopy openness in each of five phylogenetically distinct drainages, representing two major drainage systems. This sampling design helps to control for
phylogenetic effects to the extent that sites within one
drainage are closer to each other genetically than sites in
different drainages, as would be expected from the dispersal
mode of guppies. The five drainages (Fig. 1) occur in two
drainage systems, the Oropuche (the four Quare streams),
which empties into the Atlantic Ocean, and the North Slope
(the remaining six streams), which empties into the
Caribbean Sea. Previous studies have demonstrated that
the populations in these two drainage systems are genetically distinct (reviewed by Houde 1997, p. 15).
Behav Ecol Sociobiol (2007) 61:689–701
691
Fig. 1 Field sites in the Northern Range of Trinidad. The drainages,
site names, lab experiment numbers, food availability levels, and
Universal Transverse Mercator Grid coordinates (Zone 20) are as
follows: (1) Marianne drainage: Marianne River (Experiment I, high
food availability, PS 858 895); Marianne Tributary (Experiment I, low
food availability, PS 842 894). (2) Paria drainage: Paria River
(Experiment I, high food availability, PS 911 920); Paria Tributary
(Experiment I, low food availability, PS 895 907). (3) Madamas
drainage: Aqui River (Experiment II, high food availability, PS 939
887); Madamas Tributary (Experiment II, low food availability, PS
950 880). (4) Upper Quare drainage: Small Crayfish River (Experiment II, high food availability, PS 965 835); Large Crayfish River
(Experiment II, low food availability; 965 832). (5) Lower Quare
drainage: Quare 1 (Experiment I, high food availability, PS 970 806);
Quare 2 (Experiment I, low food availability, PS 969 809). The map is
modified from Houde (1997) with permission from A. E. Houde.
Field behavior observations
LI-189) over the portion of the stream where the focal
observation took place and noted the number of mature
males and females within approximately 1 m surrounding
the focal male during the majority of the focal observation
(based on visual observation). We observed 1–5 focal
males per pool, 6–26 pools per site, and 58–145 males per
site (964 focal males in total).
Focal observations were conducted in April–May 2000 at
three times of day: early (0700 to 1000), mid (1000 to
1400), and late (1400 to 1800). Food availability differences between streams persist throughout the year (Grether
et al. 2001); however, we conducted observations during
the dry season so that we could follow individual fish
without interference from rain. Focal males were selected
haphazardly and distinguished from others based on readily
visible color patterns. Observations were made by sitting
quietly next to the stream and tape recording behaviors for
5 min or until visual contact with the male was lost (male
guppies usually display more than once per minute; see
Magurran and Seghers 1994 for similar methods). We
recorded the light level [photosynthetically active radiation,
log10(μmol m−2 s−1)] using a quantum radiometer (Licor
Lab behavior observations
We performed two types of lab behavior experiments,
which we refer to as the “common garden” and “diet”
experiments. The common garden experiment involved the
same ten sites that were studied in the field and was
designed to determine whether variation between site types
(low- versus high-food-availability sites) is due to phenotypic plasticity or to fixed genetic differences in response to
692
food availability. Some of the data for this lab study were
obtained from Kolluru and Grether (2005; high food level
males). All of the males in this study were raised on either
the high food level from Kolluru and Grether (2005) or on
ad libitum food, and the details of this study are given
below. The diet experiment consisted of reanalysis of data
originally published in Kolluru and Grether (2005) and was
designed to determine whether males from low- and highfood-availability sites differ in their behavioral norms of
reaction to food intake. For this study, the males from each
of four sites (two low- and two high-food-availability) were
raised on one of two food levels, low or high. Comparison
of the asymptotic sizes of wild-caught males from low- and
high-food-availability sites with males from the lab food
treatments demonstrated that the low food level is on the
low end of the range that guppies typically experience in
the wild and the high food level is in the middle of the
range (Kolluru et al. 2006). The methods for behavioral
observations were identical to those employed in the
common garden experiment.
Common garden experiment
For this experiment, we combined data from two separate
lab experiments, one with six sites (Experiment I) and one
with four sites (Experiment II). Together, these encompassed all ten of the sites for which we collected field data
(see Fig. 1). The lab populations were housed either at the
University of California, Santa Barbara (Experiment I) or at
the University of California, Los Angeles (Experiment II).
In Experiment I, the fish were housed in a shaded
greenhouse and thus exposed to indirect sunlight. In
Experiment II, the fish were housed in a windowless room.
In both labs, water temperature was maintained at 24±2°C
and a 12:12 photoperiod was simulated using mixed
incandescent and daylight spectrum fluorescent lights. To
maximize the genetic diversity of fish used in the
experiment, we obtained offspring from 15 to 35 wild
females per site. This represents a potentially much larger
number of sires, because females mate multiply in the wild
and can store sperm for up to 8 months (Winge 1937). To
prevent the guppies from eating algae, we treated the water
in the aquaria with 2-chloro-4, 6-bis-(ethylamino)-s-triazine
(Algae Destroyer, Aquarium Pharmaceuticals) and removed
visible algae regularly.
Wild-caught females and their offspring were housed as
described in Grether (2000) and Kolluru and Grether
(2005). Food amounts for males in both experiments were
adjusted to the age and density of fish in the tank. In
Experiment I, the fish were fed flake food to satiation
[detailed protocol in Grether (2000)]. In Experiment II, the
fish were fed twice daily (once daily on weekends) using a
specially designed feeding device that delivered precise
Behav Ecol Sociobiol (2007) 61:689–701
quantities of ground flake food to each tank. The food
consisted of a mixture of spray-dried white fishmeal
(41.8%), wheat flour (47%), vegetable oil (2.0%), vitamin
premix (1.0%), and gelatin (8.1%). The estimated protein
content was 40%, and the fat content was 10% (Lamon
2001, personal communication). Further details about the
feeding protocol can be found in Kolluru and Grether
(2005). The fish used for behavioral observations in
Experiment I were second generation lab-born, and fish
used in Experiment II were first generation lab-born.
After being sexed under a dissecting microscope, males
were housed in 8-l tanks at densities of 1–4 males per tank,
and females were housed in 38-l tanks at densities of 20
females per tank. Housing density did not affect male
behavior in Experiment II (Kolluru and Grether 2005). To
allow males to have courtship experience, we housed one
mature stock female in each male tank for at least 1 week
before behavioral observations. Females remained virgins
until they were used in observations.
We used an open-aquarium design in which the fish
could interact directly (Houde 1997), allowing us to
simultaneously examine aggressive, courtship, and foraging
behavior. Observations were conducted in 120-l (Experiment I) and 180-l (Experiment II) aquaria with natural,
multicolored gravel bottoms and plastic bubblers connected
to undergravel filters in windowless rooms maintained on
the same light/dark schedule as the respective labs. The
observation aquaria were covered with brown paper on
three sides, and observations were made from the fourth
side. Each aquarium was illuminated from the top with one
daylight-spectrum fluorescent tube. Otherwise, the room
was dark to maximize the visibility of the fish to the
observer and to minimize the visibility of the observer to
the fish.
To minimize the effects of competition for food
(Magurran and Seghers 1991), we fed the fish to satiation
twice per observation day and regularly removed visible
algae from the observation aquaria. We filtered the water in
the aquaria using a high-flow-rate charcoal canister filter
(Marineland Magnum 350 convertible canister filter, Moorpark, CA) after each set of observations to minimize
chemical effects on the behavior of fish in subsequent
observations. To avoid artificially inflating male–male
aggression, we used an even sex ratio (3:3), very low
densities of fish per observation tank, and males that had
not been housed together (see Houde 1997; Grether 2000).
We also minimized body size disparities within male and
female groups.
Behavior observations began within 2 h after the lights
came on and were concluded within 4 h after the lights
came on. A trial was initiated by releasing the three males
chosen for testing into the observation aquarium after their
color patterns were sketched. Males were chosen based on
Behav Ecol Sociobiol (2007) 61:689–701
body size similarities and not based on color patterns.
Females were released into the observation aquarium
shortly after the males. The fish were then fed. On the
following morning, the fish were fed again, and the first
observation session began at least 15 min after the feeding.
We performed at least three replicate focal samples of 5 min
per male (with additional replicates added if a male did not
perform courtship displays in at least two of the initial three
replicates), alternating between males in a predetermined,
random order. A minimum of 20 min elapsed between
consecutive focal samples on a given male, and fish were
fed again after the second sample (see Kolluru and Grether
2005 for additional details). We conducted lab observations
on 374 males (Marianne drainage, 72; Paria drainage, 72;
Madamas drainage, 78; Upper Quare drainage, 69; Lower
Quare drainage, 83) and an equal number of females.
Immediately following their use in observations, males and
females were weighed to the nearest 0.1 mg, and their
standard length was measured using digital calipers to the
nearest 0.01 mm.
Behavior variables
We recorded the following variables: time spent foraging,
competing (two or more males simultaneously following or
displaying to the same female) and swimming (including
time spent searching for mates), courtship display rate,
sneak copulation rate (forced copulation attempts not
preceded by a display, in which gonopodial contact with
the female’s ventral surface was visible), escalated competition rate (competitions including chases and/or bites
between males), dominance interaction rate (supplanting,
displaying, chasing, or biting directed from one male to
another while neither was following or courting a female),
and escalated dominance interaction rate (dominance
interactions including chases and/or bites between males).
Table 1 PCA with varimax
rotation of male P. reticulata
behavior measured in the field
(N=925)
Loadings>|0.50| are shown in
bold.
Variable
693
Dominance interactions were usually distinctly one-sided,
and thus one male could be classified as dominant and the
other as subordinate. These measures of escalated competition and dominance are slightly different from those in
Kolluru and Grether (2005). In that study, we reported the
proportion of competition and dominance interactions that
included escalations, rather than the frequency of escalated
interactions, as we do here. Including the proportion
variables would have resulted in too many missing values
to perform a meaningful Principal Components Analysis
(PCA; see below) because not all males were observed in
competitive or dominance interactions.
Data analysis
All data were transformed to meet parametric assumptions,
and all analyses were performed using JMP 3.2.2 (SAS
Institute, Cary, NC). To extract independent components of
variation from the field data, we performed PCA; Tabachnick
and Fidell 2001). The field PCA yielded three components
with eigenvalues >1, which explained 66.4% of the variation
in behavior (Table 1). We interpreted the components based
on the loadings (correlations between components and the
original variables; Tabachnick and Fidell 2001) with
absolute values greater than 0.5 (similar methods have
been used in other studies; e.g., Reyer et al. 1998; Zuk et al.
1998). A varimax rotation, which maximizes high loadings
and minimizes low loadings (Tabachnick and Fidell 2001),
yielded components interpretable as “courtship vs foraging”, “dominance”, and “competition vs searching for
mates”. To use the field components of behavior to analyze
the lab data, we first validated the approach by performing
an independent PCA of the lab data, entering the same
variables as for the field PCA. This PCA yielded four
components with eigenvalues >1.0, that explained a total of
73.4% of the variation in the original data (Table 2). The
Component
1 (Courtship vs foraging)
2 (Dominance)
3 (Competition vs
searching for mates)
Time foraging
Time following females
Time competing
Time searching for mates
Courtship display rate
Sneak copulation rate
Dominance interaction rate
Escalated competition rate
Escalated dominance rate
−0.772
0.772
0.263
−0.244
0.734
0.320
0.038
−0.078
0.025
0.086
0.174
0.008
−0.211
0.096
−0.149
−0.963
−0.141
−0.960
−0.117
0.475
0.789
−0.713
0.258
−0.138
0.017
0.749
−0.033
Eigenvalue
Explained variance (%)
2.97
32.95
1.95
21.63
1.07
11.86
694
Behav Ecol Sociobiol (2007) 61:689–701
Table 2 PCA with varimax rotation of male P. reticulata behavior measured in the lab (N=374)
Variable
Component
1 (Dominance+
escalated competition)
2 (Courtship+
competition vs foraging)
3 (Courtship vs searching
for mates)
4 (Sneak
copulations)
Time foraging
Time following females
Time competing
Time searching for mates
Courtship display rate
Sneak copulation rate
Dominance interaction rate
Escalated competition rate
Escalated dominance rate
−0.001
0.18
−0.18
−0.003
−0.03
0.02
−0.92
−0.75
−0.94
−0.58
0.14
0.85
−0.04
0.71
0.007
0.07
0.18
−0.06
0.37
−0.80
0.21
0.79
−0.37
−0.03
0.072
0.03
0.086
0.25
0.15
0.04
0.08
0.13
0.95
0.03
−0.06
−0.008
Eigenvalue
Explained variance (%)
2.53
28.06
2.04
22.63
1.03
11.48
1.01
11.23
Loadings>|0.50| are shown in bold.
lab PCA yielded similar components as those we obtained
for the field, with the exception that a fourth axis consisting
solely of sneak copulation rate emerged. Sneak copulation
rate did not vary with site food availability in the lab
(F(1, 249) =3.86, P=0.10) or with food availability in the
field (F(1, 954) = 2.42, P= 0.17). Given the similarity
between the lab and field data sets, we used the
components obtained from the field PCA to construct
component scores separately for the field and lab datasets
as described below (see Rotenberry and Wiens 1998 for
similar methods). To standardize among rate and duration
variables, we converted all data to z scores (with a mean of
zero and standard deviation of one); we then constructed
component scores using unit (±1) component score
coefficients for the variables with loadings >0.5 (Table 1)
and used the component scores for subsequent analyses.
We switched the signs of the coefficients for Component 2
(dominance) because the component had negative loadings.
To determine whether low- and high-food-availability
sites differed in mating strategies, we performed analyses of
variance (ANOVAs) on the field component scores,
employing models with a random-effects site term nested
within drainage system and site food availability and fixedeffects site food availability and drainage system terms. To
examine the influence of time of day, male and female
density, sex ratio [m/(m+f)], and light level, we entered
each of these covariates into the ANOVA model separately,
correcting for multiple tests as noted in the text and tables.
We standardized data from experiments I and II of the
common garden experiment to control for differences
between them in factors such as the mean age of males
and light conditions. This standardization could not have
biased the results because of the way in which we carried
out the ANOVA; that is, because sites were nested within
drainage and different drainages were included in the two
lab studies. To construct lab component scores, we
performed an ANOVA with experiment (I or II) as the sole
factor, taking residuals on the experiment to standardize
between lab experiments, and added the overall mean (both
experiments combined) to each residual to make the results
more comparable to field values. We then constructed the
component scores as described above. The subsequent
ANOVA models included a random-effects site term nested
within drainage system and food availability and a randomeffects male group (the group of three males tested
together) term. The model also included fixed-effects site
food availability and drainage system terms. Analyses of
covariance including male standard length (distance from
Table 3 Analysis of variation in male guppy mating behavior in the field as a function of site, drainage system, and site food availability
Component 1
(courtship vs foraging)
Site (drainage system, site food availability)
Drainage system
Site food availability
Drainage system×site food availability
Values are F(df); P.
4.35(6,
0.20(1,
2.76(1,
2.49(1,
952);
0.0002
0.67
;
952) 0.15
952); 0.17
952);
Component 2 (dominance)
14.30(6,
1.12(1,
1.06(1,
0.30(1,
917);
<0.0001
0.33
;
917) 0.34
917); 0.60
917);
Component 3 (competition vs
searching for mates)
3.26(6,
10.14(1,
8.17(1,
0.16(1,
954);
0.004
0.019
;
954) 0.028
954); 0.70
954);
Behav Ecol Sociobiol (2007) 61:689–701
high food availability sites
695
low food availability sites
Mean component score
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
Courtship vs foraging
Dominance
Competition vs
searching for mates
Fig. 2 Mean component scores for male guppies (P. reticulata) from
field observations. Bars show square-root transformed, least-squared
means±SE
the lower jaw to the caudal peduncle) as a covariate
revealed no significant relationship between length and the
component scores (all P>0.24), so length was not included
in the final models.
For the diet experiment, we reanalyzed data from
Kolluru and Grether (2005) by constructing the component
scores as described above and performing subsequent
ANOVAs employing a model with site food availability
and lab food level as main effects. The model also included
two random-effects terms: site nested within site food
availability and male group (the group of three males tested
together) nested within site and food level.
Results
Field observations
The ten sites varied in mean levels of all three components
(Table 3). Males in high-food-availability sites had higher
scores for Component 3 (competition vs searching for
mates) than males in low-food-availability sites (Table 3;
Fig. 2), indicating that males in the high-food-availability
sites invested relatively more effort into aggressively
Table 4 Variation in environmental factors among the ten
sites sampled during field focal
observations
Male and female densities are
square-root transformed
numbers of fish, sex ratio is
m/(m+f), and light level is
log10(μmol m−2 s−1 ). Values
are mean±SE.
Site
High food availability
Aqui River
Marianne River
Paria River
Small Crayfish River
Quare 1
Low food availability
Madamas Tributary
Marianne Tributary
Paria Tributary
Large Crayfish River
Quare 2
competing for females. There were no differences between
low- and high-food-availability sites for the other two
components. The effort that males invested into courtship at
the expense of foraging (Component 1) varied with time of
day; scores for this component were greater early (leastsquares mean±SE=0.20±0.17) and late in the day (0.21±
0.14) than at midday (−0.26±0.12; F(2, 944) =3.98, P=
0.019). The other components did not vary with time of day,
and none of the interactions involving time of day was
significant (all P>0.06).
The ten sites differed with respect to male and female
density, sex ratio, and light levels during the focal
observations (Tables 4, 5). Low-food-availability sites had
lower light levels; no other covariates differed between low
and high-food-availability sites. We found a positive correlation between population mean scores on Component 1
(courtship vs foraging) and site mean sex ratio (N=10,
Spearman rho=0.75, P=0.013) and a negative correlation
between site mean scores on Component 3 (dominance) and
site mean density of females (N=10, Spearman rho=−0.65,
P=0.043), but none of these correlations was significant
after correcting for multiple tests (control of the false
discovery rate; Verhoeven et al. 2005; k=9 correlations).
The other correlations between site mean component scores
and light level, sex ratio, and male and female density were
not statistically significant (all P>0.07). We also examined
the correlations between site mean component scores in the
lab and in the field and found none of these to be statistically
significant (all |Spearman rho|<0.54, all P>0.10).
Addition of each of the four covariates (male and female
density, sex ratio, and light level) to the ANOVA model
separately did not substantially alter the contribution of the
main effects (site, drainage system, and food availability) to
variation in male behavior. Component 1 (courtship vs
foraging) decreased with light level (standardized β=−0.16,
F(1, 535) =12.61, P=0.0004), Component 2 (dominance)
increased with female density (standardized β = 0.10,
F(1, 820) =6.39, P=0.012) and decreased with sex ratio
(standardized β=−0.14, F(1, 815) =17.30, P<0.0001), and
Male density
Female density
Sex ratio
Light level
1.76±0.08
1.64±0.08
1.62±0.09
2.55±0.07
1.92±0.10
2.15±0.08
1.93±0.09
1.83±0.10
3.29±0.07
2.88±0.11
0.39±0.02
0.43±0.02
0.43±0.02
0.39±0.01
0.31±0.02
1.76±0.06
1.44±0.06
1.53±0.07
1.46±0.07
1.33±0.09
1.55±0.07
1.99±0.07
1.41±0.09
1.83±0.08
1.75±0.08
2.61±0.08
2.30±0.08
1.90±0.10
3.13±0.08
2.16±0.09
0.28±0.02
0.43±0.02
0.36±0.02
0.27±0.02
0.40±0.02
1.25±0.07
1.19±0.06
1.23±0.06
1.34±0.06
0.97±0.16
696
Behav Ecol Sociobiol (2007) 61:689–701
Table 5 Analysis of variation in male and female density, sex ratio, and light level during focal observations as a function of site, drainage
system, and food availability
Male density
Site
Drainage system
Site food availability
Drainage system×site food availability
9.89(6,
4.57(1,
2.00(1,
1.63(1,
856);
<0.0001
0.08
856); 0.21
856); 0.25
856);
Female density
Sex ratio
Light level
19.58(6,
9.01(1,
0.08(1,
2.17(1,
14.50(6,
1.03(1,
0.82(1,
0.31(1,
3.43(1,
1.82(1,
10.46(1,
0.35(1,
857);
<0.0001
0.023
857); 0.79
857); 0.19
857);
852);
<0.0001
0.35
852); 0.40
852); 0.60
852);
537);
0.003
0.21
537); 0.011
537); 0.57
537);
Values are F(df); P.
Component 3 (competition vs searching for mates) increased with male density (standardized β=0.20, F(1, 855) =
32.04, P<0.0001), female density (standardized β=0.12,
F(1, 856) =9.75, P=0.002), and sex ratio (standardized
β=0.13, F(1, 851) =13.59, P=0.0002). All of these results
remained significant after a Bonferroni correction within
each component (i.e., for four tests; α=0.0125).
males from high-food-availability sites. In other words, the
norm of reaction with food intake was greater for males
from low-food-availability sites. For the dominance component, the norm of reaction with food intake was greater
for males from high-food-availability sites.
Discussion
Common garden experiment
All three behavioral factors varied significantly among
male groups (all P<0.017), and factors 2 (F(1, 354) =8.28,
P<0.0001) and 3 (F(1, 354) =8.57, P<0.0001) varied among
sites. None of the factors varied significantly with site food
availability (factor 1: F(1, 354) =0.02, P=0.89; factor 2:
F(1, 354) =4.32, P=0.06; factor 3: F(1, 354) =1.01, P=0.34),
and none of the other terms in any of the models was
significant (all P>0.06). We performed separate analyses of
the two experiments to ensure that the results were not a
spurious consequence of combining the two datasets, and
found almost identical results; none of the factors varied
with site food availability.
Diet experiment
Males raised on the high food level scored higher on
component 2 (dominance) than did males raised on the low
food level (Table 6; Fig. 3a–c). All three components
showed significant site food availability×lab food level
interactions; for components 1 and 3, this was caused by
greater differences between low- and high-food-level
groups for males from low-food-availability sites than for
We found three independent components of variation in male
behavior in the field: (1) courtship versus foraging, (2)
dominance interactions, and (3) interference competition
versus searching for mates (Table 1; also see Croft et al.
2003). Each of these behavioral components varied among
sites but only the third correlated with food availability in the
field. Compared with males at low-food-availability sites,
males at high-food-availability sites devoted more effort to
interference competition (aggressively competing over
females) versus searching for mates. This difference between
high- and low-food-availability populations disappeared in
the common garden experiment, which suggests that it
primarily resulted from phenotypic plasticity and not genetic
divergence between sites (also see Carroll and Corneli 1999;
Foster and Endler 1999). The diet experiment confirmed that
food intake increases interference competition relative to
mate searching, but this was only true for fish from lowfood-availability sites. Food intake also had a direct positive
effect on the rate and intensity of dominance interactions,
and the strength of this effect was greater in fish from highfood-availability sites. These results generally support the
prediction that males engage in less energetically demanding
mating tactics when food is scarce.
Table 6 Analysis of variation in male guppy mating behavior in the lab as a function of site food availability and lab food level
Component 1
(courtship vs foraging)
Site
Lab
Site
Site
(site food availability)
food level
food availability
food availability×lab food level
11.40(2,
0.81(1,
1.27(1,
4.98(1,
202);
<0.0001
0.37
202); 0.38
202); 0.027
202);
The male group effect is not shown. Values are F(df); P.
Component 2 (dominance)
30.98(2,
11.55(1,
0.007(1,
15.06(1,
202);
<0.0001
0.001
202); 0.94
202); 0.0001
202);
Component 3 (competition vs
searching for mates)
12.12(2,
0.70(1,
1.09(1,
7.56(1,
202);
<0.0001
0.40
202); 0.41
202); 0.007
202);
Behav Ecol Sociobiol (2007) 61:689–701
low food
697
high food
1.5
a) courtship vs foraging
1
0.5
0
-0.5
-1
-1.5
low food availability
high food availability
1.5
b) dominance
1
0.5
0
-0.5
-1
-1.5
low food availability
high food availability
1.5
c) competition vs searching for mates
1
0.5
0
-0.5
-1
-1.5
low food availability
high food availability
Site food availability
Fig. 3 Mean component scores for Component 1 (courtship versus
foraging; a), Component 2 (dominance; b), and Component 3
(competition versus searching for mates; c) for male guppies (P.
reticulata) from either low- or high-food-availability sites, raised on
either low or high food levels. Data are reanalyzed from Kolluru and
Grether (2005). Bars show square-root transformed, least-squared
means±SE
Presumably, the expected payoffs of the different mating
tactics exhibited by male guppies are frequency dependent.
Our results suggest that the relative success of the
interference competition tactic increases as food availability
increases, as would be expected if this tactic is more
energy-demanding than the alternative tactic of searching
for uncourted females. One surprising discrepancy between
our field and laboratory results is that dominance interactions were strongly affected by food intake in the
laboratory but did not correlate with food availability in
the field. It seems likely that food intake affects dominance
interactions in the field but that this effect was obscured by
countervailing environmental or genetic factors. The common garden experiment showed that populations vary
genetically along this behavioral axis, independent of food
availability. This could account for the lack of a food
availability effect in the field study. Whether better-fed
males invest more effort in dominance interactions in the
field is an open question.
Other factors besides food intake could potentially be
responsible for the association between food availability
and interference competition in the field. Ambient light
levels and male size at maturity both covary with food
availability in the field (Grether et al. 1999, 2001; this
study) and could directly affect male behavior. Male size
has been suggested to influence courtship and sneak
copulation rates (Rodd and Sokolowski 1995; Magellan et
al. 2005). Nevertheless, interference competition was not
correlated with male size within food level groups in our
diet experiment (Kolluru and Grether 2005) or with light
levels during focal observations in our field study.
Other studies support a role for food availability in the
development or expression of guppy mating tactics. Fraser
et al. (2004) found that denying male guppies the
opportunity to feed at night reduced their diurnal courtship
activity (time spent following females) and suggested that
release from predation allows males to forage more at night
and thereby court more during the day. Although these
authors did not examine male–male interactions or courtship displays, their results suggest that costly mating
activity is limited by food intake. In lab studies, hungry
male guppies feed before courting females (Abrahams
1993), and hungrier males spend more time foraging than
more satiated males, even in the presence of females
(Griffiths 1996).
Genetic divergence among sites need not be in the mean
value of behaviors but may instead take the form of
genotype by environment interactions, so that the degree of
response to food intake varies genetically among sites
(Rodd and Sokolowski 1995; Carroll and Corneli 1999;
Thompson 1999; Hughes et al. 2005). Indeed, comparing
mean differences among sites in the lab under ad libitum
food conditions may not reveal genetic differences among
sites if the differences are in the norm of reaction (Weitere
et al. 2004). Our diet experiment, which included guppies
from two low- and two high-food-availability sites,
revealed that all three behavioral components showed
significant interactions between site food availability and
lab food level, suggesting that low- and high-foodavailability sites differ in their behavioral norms of reaction
to food intake. For the courtship-versus-foraging and
competition-versus-searching-for-mates components, the
norms of reaction were steeper for males from low-food-
698
availability sites (Fig. 3). This may be due to greater
variability in food availability in low-food-availability sites
(see below), which is expected to favor greater flexibility
(reviewed by Komers 1997). If the benefits of behavioral
plasticity are reduced in high-food-availability sites, for
example because retaining the ability to assess food intake
is not necessary when food availability is consistently high,
then plasticity is expected to be reduced (Komers 1997;
DeWitt et al. 1998). Similar variation in the degree of
plasticity among guppy populations was described by Rodd
and Sokolowski (1995) and Rodd et al. (1997): males from
high-predation sites were less plastic in their mating
behavior and in some life history traits than males from
low-predation sites. Interestingly, low-predation sites usually have lower food availability than high-predation sites
(Reznick et al. 2001), so that the greater flexibility in these
sites may be due to food availability differences rather than
to predation level differences (in our study, predation and
food availability were not confounded).
What has prevented the mean levels of male tactics from
diverging genetically along the food availability gradient?
A logical explanation is that gene flow along the gradient
has prevented local adaptation. This explanation seems
unlikely, however, because the sites included in this study
are separated by multiple dispersal barriers and have
diverged genetically in female mate preferences (Grether
2000; Grether et al. 2005) and male coloration (Grether
2000; Grether and Kolluru, unpublished data). Moreover,
we did find evidence for genetic differentiation in behavior
among sites in the current study (note significant site terms
in Table 6; also see Kolluru and Grether 2005). The
explanation we favor is that the plastic responses of male
mating tactics to food intake (as revealed by the diet
experiment; Kolluru and Grether 2005) are evolved norms
of reaction that reduce divergent selection along the food
availability gradient. For this explanation to make sense,
food availability would have to vary sufficiently within
sites for selection to maintain the norms of reaction. Food
availability does vary between pools within streams
(Grether et al. 2001) and can change rapidly when trees
fall and create temporary gaps in the forest canopy (Grether
and Kolluru, unpublished data). Floods may also cause
transient changes in food availability either by reducing the
densities of guppies or reducing algal standing crops
(Grether et al. 2001). Thus, it seems plausible that this is
a case of adaptive plasticity reducing genetic divergence
between populations. Our finding that two of the three
behavioral components (courtship versus foraging and
competition versus mate searching) were more strongly
influenced by food intake in fish derived from low-foodavailability sites is consistent with the hypothesis that the
reaction norms are maintained by selection, to the extent
that food availability is more stable (or above some
Behav Ecol Sociobiol (2007) 61:689–701
energetic threshold) at high-food-availability sites. Tree
falls have a larger (positive) effect on algae production in
low-food-availability streams (our personal observation and
unpublished data). Floods tend to have a larger impact on
high-food-availability streams, but because algal standing
crops and guppy densities are both reduced, food availability is not greatly affected by floods (Grether et al.
2001).
Light conditions are known to influence guppy reproductive behavior (Long and Rosenqvist 1998; Gamble et al.
2003), presumably because increasing light levels increase
the perceived risk of predation (reviewed by Houde 1997).
In our study, courtship occurred at a higher rate early and
late in the day than at midday and decreased with light level
within sites, indicating that males perform courtship displays when they are less conspicuous to potential predators,
as first described by Endler (1987) using guppies in
artificial streams. This pattern likely occurs because guppy
color patterns are most conspicuous to predators at midday
and under high light conditions (Endler 1987). Although
our sites are subject to low fish predation intensity, they
contain R. hartii and freshwater prawns (Macrobrachium
crenulatum; Endler 1978; Rodd and Reznick 1991; Millar
et al. 2006) and may be at risk from visually orienting aerial
predators (Endler 1978; Templeton and Shriner 2004).
In many species, the success of mating tactics is
influenced by male density and operational sex ratio
(Carroll 1993; Quinn et al. 1996; Mills and Reynolds
2003; Lodé et al. 2004), and aggressive mating tactics may
be employed more frequently when the sex ratio is more
male biased (reviewed by Jirotkul 1999a). In guppies,
sneak copulations increase (Evans and Magurran 1999; but
see Jirotkul 1999a), courtship displays decrease (Jirotkul
1999a; Evans and Magurran 1999), and male-male competition increases (Jirotkul 1999a; Price and Rodd (2006) as
the sex ratio becomes more male biased. Employing much
higher densities than those we observed in the field, Jirotkul
(1999b) found that aggressive competition also increased
with male density in the lab.
Consistent with these observations, we found that males
spent more time competing when sex ratios were more male
biased and with increasing male density. In contrast,
dominance interactions appeared to be more common when
competition for females was least intense. Within sites,
dominance interactions increased with female density and
decreased with sex ratio, suggesting that it may be more
profitable for males to spend time establishing and
maintaining dominance hierarchies when competition for
females is less intense. These results also emphasize that
competition and dominance probably serve different purposes for male guppies (Kolluru and Grether 2005).
Dominance interactions, which involve (sometimes prolonged) fights between males away from females, are likely
Behav Ecol Sociobiol (2007) 61:689–701
to be involved in establishing long-term dominance
hierarchies among males and may only occur when females
are abundant and males can afford to spend time in
prolonged interactions with each other. Competition, which
involves directly fighting to acquire or maintain access to
females, appears instead to be important in the shorter term
and may be a tactic better employed when competition for
females is intense.
The sex ratios we observed were consistent with those
obtained by Pettersson et al. (2004). Their snapshot survey
of 11 guppy populations revealed that sex ratios departed
significantly from 50:50 in most cases, and that most sites
were female biased. It is important to note that because we
did not determine the receptivity of females in the field, we
could not measure the true operational sex ratio (Emlen and
Oring 1977). Most female guppies are not sexually
receptive at any given time; fewer than 10% respond
positively to male displays (Magurran and Seghers 1994).
Therefore, it is likely that we overestimated the degree to
which operational sex ratios were female biased.
Whether male–male competition is important for
guppies, and indeed whether it even occurs in the wild,
has been the subject of debate (Farr 1975, 1989; Luyten
and Liley 1991; Kodric-Brown 1992, 1993; Houde 1997).
Recent lab studies have confirmed that males employ
aggressive mating tactics (Jirotkul 1999a,b; Kelly and
Godin 2001; Kolluru and Grether 2005; Price and Rodd
2006) and suggest that more aggressive males may enjoy
higher mating success (Kodric-Brown 1993; Price and Rodd
2006). We commonly observed aggression in the context
of mating in the field at sites ranging from small, isolated
pools with high guppy densities to larger streams with fastflowing currents and low guppy densities. Although males
move among pools (Croft et al. 2003), we sometimes
resighted males on multiple visits. At the Marianne
Tributary (see Fig. 1 legend), we witnessed rudimentary
territorial defense of an area of the stream, a behavior
commonly seen in other poeciliid species (reviewed by Farr
1989) and in foraging guppies (Magurran and Seghers
1991). Male guppies remember the individual identities of
others and use this information in social contexts (Dugatkin
and Sargent 1994; Dosen and Montgomerie 2004). Dominance hierarchies based on prior information gained by
males about each other (e.g., Earley et al. 2003), which
may be necessary for aggression to influence reproductive
success (Farr 1989), may therefore occur in some
populations. Recently, Price and Rodd (2006) demonstrated that male guppies unfamiliar with each other are
more aggressive than familiar males, presumably because
unfamiliar males are in the process of establishing
dominance relationships.
Our study emphasizes that variation in guppy mating
behavior along food availability gradients is a complex
699
mixture of behavioral plasticity, genetic divergence, and
genotype by environment interactions (see also Rodd and
Sokolowski 1995; Houde 1997). Phenotypic plasticity such
as that we have observed is important because it may
reduce selection along food availability gradients (Price et
al. 2003; West-Eberhard 1989). In addition, the behavioral
differences we describe between low- and high-foodavailability sites are similar to those found by some authors
comparing low- and high-predation sites (Endler 1995;
Houde 1997). Because high-predation localities tend to be
more open, with higher light levels and greater primary
productivity, than low-predation localities (Reznick et al.
2001; Grether et al. 2001), at least some of the behavioral
variation along predation gradients may be attributable to
concomitant variation in food availability (Magurran and
Seghers 1994).
Acknowledgment We are grateful to Randy Tashjian and Wendy
Mayea for helping to transcribe data from audiotapes, and C. St. Mary
and two anonymous reviewers for their comments on an earlier
version of the manuscript. In Trinidad, we thank the Sinanan family
for housing accommodations, the Ministry of Food Production,
Marine Exploitation, Forestry, and the Environment for permits to
collect guppies, and the Water and Sewage Authority for permission to
work in the Quare drainages. This study was supported by National
Science Foundation grants IBN-0001309 to GFG and IBN-0130893 to
GFG and GRK. The experiments comply with current US laws.
References
Abrahams MV (1993) The trade-off between foraging and courting in
male guppies. Anim Behav 45:673–681
Belovsky GE, Slade JB, Chase JM (1996) Mating strategies based on
foraging ability: an experiment with grasshoppers. Behav Ecol
7:438–444
Blanckenhorn WU, Preziosi RF, Fairbairn DF (1995) Time and energy
constraints and the evolution of sexual size dimorphism—to eat
or to mate? Evol Ecol 9:369–381
Brockmann HJ (2001) The evolution of alternative strategies and
tactics. Adv Study Behav 30:1–51
Carroll SP (1993) Divergence in male mating tactics between two
populations of the soapberry bug: I. guarding versus nonguarding.
Behav Ecol 4:156–164
Carroll SP, Corneli PS (1995) Divergence in male mating tactics
between two populations of the soapberry bug: II. genetic change
and the evolution of a plastic reaction norm in a variable social
environment. Behav Ecol 6:46–56
Carroll SP, Corneli PS (1999) The evolution of behavioral norms of
reaction as a problem in ecological genetics: theory, methods, and
data. In: Foster SA, Endler JA (eds) Geographic variation in
behavior. Oxford University Press, New York, pp 52–68
Croft DP, Albanese B, Arrowsmith BJ, Botham M, Webster M,
Krause J (2003) Sex-biased movement in the guppy (Poecilia
reticulata). Oecologia 137:62–68
DeWitt TJ, Sih A, Wilson DS (1998) Costs and limits of phenotypic
plasticity. Trends Ecol Evol 13:77–81
Dosen LD, Montgomerie R (2004) Mate preferences by male guppies
(Poecilia reticulata) in relation to the risk of sperm competition.
Behav Ecol Sociobiol 55:266–271
700
Dugatkin LA, Sargent RC (1994) Male–male association patterns and
female proximity in the guppy, Poecilia reticulata. Behav Ecol
Sociobiol 35:141–145
Dussault GV, Kramer DL (1981) Food and feeding behavior of the guppy,
Poecilia reticulata (Pisces: Poeciliidae). Can J Zool 59:684–701
Earley RL, Tinsley M, Dugatkin LA (2003) To see or not to see: does
previewing a future opponent affect the contest behavior of green
swordtail males (Xiphophorus helleri)? Naturwissenschaften
90:226–230
Emlen ST, Oring LW (1977) Ecology, sexual selection, and the
evolution of mating systems. Science 197:215–223
Endler JA (1978) A predator’s view of animal color patterns. Evol
Biol 11:319–364
Endler JA (1987) Predation, light intensity and courtship behavior in
Poecilia reticulata (Pisces: Poeciliidae). Anim Behav 35:1376–1385
Endler JA (1995) Multiple-trait coevolution and environmental
gradients in guppies. Trends Ecol Evol 10:22–29
Evans JP, Magurran AE (1999) Male mating behavior and sperm
production characteristics under varying sperm competition risk
in guppies. Anim Behav 58:1001–1006
Farr JA (1975) The role of predation in the evolution of social
behavior of natural populations of the guppy, Poecilia reticulata
(Pisces: Poeciliidae). Evolution 29:151–158
Farr JA (1989) Sexual selection and secondary sexual differentiation
in poeciliids: determinants of male mating success and the
evolution of female choice. In: Meffe GK, Snelson FF (eds)
Ecology and evolution of livebearing fishes (Poeciliidae).
Prentice Hall, Englewood Cliffs, NJ, pp 91–123
Foster SA, Endler JA (1999) Thoughts on geographic variation in
behavior. In: Foster SA, Endler JA (eds) Geographic variation in
behavior. Oxford University Press, New York, pp 287–305
Fraser DF, Gilliam JF, Akkara JT, Albanese BW, Snider SB (2004)
Night feeding by guppies under predator release: effects on
growth and daytime courtship. Ecology 85:312–319
Gamble S, Lindholm AK, Endler JA, Brooks R (2003) Environmental
variation an the maintenance of polymorphism: the effect of
ambient light spectrum on mating behavior and sexual selection
in guppies. Ecol Lett 6:463–472
Grether GF (2000) Carotenoid limitation and mate preference
evolution: a test of the indicator hypothesis in guppies (Poecilia
reticulata). Evolution 54:1712–1724
Grether GF, Hudon J, Millie DF (1999) Carotenoid limitation of
sexual coloration along an environmental gradient in guppies.
Proc R Soc Lond B 266:1317–1322
Grether GF, Millie DF, Bryant MJ, Reznick DN, Mayea W (2001)
Rain forest canopy cover, resource availability, and life history
evolution in guppies. Ecology 82:1546–1559
Grether GF, Kolluru GR, Rodd FH, de la Cerda J, Shimazaki K
(2005) Carotenoid availability affects the development of a
colour-based mate preference and the sensory bias to which it is
genetically linked. Proc R Soc Lond B 272:2181–2188
Griffiths SW (1996) Sex differences in the trade-off between feeding
and mating in the guppy. J Fish Biol 48:891–898
Gross MR (1996) Alternative reproductive strategies and tactics:
diversity within sexes. Trends Ecol Evol 11:92–98
Houde AE (1997) Sex, color, and mate choice in guppies. Princeton
University Press, New Jersey
Houde AE, Endler JA (1990) Correlated evolution of female mating
preferences and male color patterns in the guppy Poecilia
reticulata. Science 248:1405–1408
Hughes KA, Rodd FH, Reznick DN (2005) Genetic and environmental effects on secondary sex traits in guppies (Poecilia reticulata).
J Evol Biol 18:35–45
Jirotkul M (1999a) Operational sex ratio influences female preference
and male–male competition in guppies. Anim Behav 58:287–
294
Behav Ecol Sociobiol (2007) 61:689–701
Jirotkul M (1999b) Population density influences male–male competition in guppies. Anim Behav 58:1169–1175
Jirotkul M (2000) Male trait distribution determined alternative
mating tactics in guppies. J Fish Biol 56:1427–1434
Kelly CD, Godin J-GJ (2001) Predation risk reduces male–male
sexual competition in the Trinidadian guppy (Poecilia reticulata). Behav Ecol Sociobiol 51:95–100
Kodric-Brown A (1992) Male dominance can enhance mating success
in guppies. Anim Behav 44:165–167
Kodric-Brown A (1993) Female choice of multiple male criteria in
guppies: interacting effects of dominance, coloration and courtship. Behav Ecol Sociobiol 32:415–420
Kolluru GR, Grether GF (2005) The effects of resource availability on
alternative mating tactics in guppies (Poecilia reticulata). Behav
Ecol 16:294–300
Kolluru GR, Grether GF, South SH, Dunlop E, Cardinali A, Liu L,
Carapiet A (2006) The effects of carotenoid and food availability
on resistance to a naturally occurring parasite (Gyrodactylus
turnbulli) in guppies (Poecilia reticulata). Biol J Linn Soc
89:301–309
Komers PE (1997) Behavioral plasticity in variable environments.
Can J Zool 75:161–169
Lodé T, Holveck M-J, Lesbarreres D, Pagano A (2004) Sex-biased
predation by polecats influences the mating system of frogs. Proc
R Soc Lond B 271:S399–S401
Long KD, Rosenqvist G (1998) Changes in male guppy courting
distance in response to a fluctuating light environment. Behav
Ecol Sociobiol 44:77–83
Luyten PH, Liley NR (1991) Sexual selection and competitive mating
success of male guppies (Poecilia reticulata) from four Trinidad
populations. Behav Ecol Sociobiol 28:329–336
Magellan K, Pettersson LB, Magurran AE (2005) Quantifying male
attractiveness and mating behavior through phenotypic size
manipulation in the Trinidadian guppy, Poecilia reticulata.
Behav Ecol Sociobiol 58:366–374
Magurran AE, Seghers BH (1991) Variation in schooling and
aggression amongst guppy (Poecilia reticulata) populations in
Trinidad. Behaviour 118:214–234
Magurran AE, Seghers BH (1994) Sexual conflict as a consequence of
ecology: evidence from guppy, Poecilia reticulata, populations in
Trinidad. Proc R Soc Lond B 225:31–36
Millar NP, Reznick DN, Kinnison MT, Hendry AP (2006) Disentangling the selective factors that act on male color in wild
guppies. Oikos 113:1–12
Mills SC, Reynolds JD (2003) Operational sex ratio and alternative
reproductive behaviors in the European bitterling, Rhodeus
sericeus. Behav Ecol Sociobiol 54:98–104
Moczek AP, Emlen DJ (1999) Proximate determination of male horn
dimorphism in the beetle Onthophagus taurus (Coleoptera:
Scarabaeidae). J Evol Biol 12:27–37
Pettersson LB, Ramnarine IW, Becher SA, Mahabir R, Magurran AE
(2004) Sex ratio dynamics and fluctuating selection pressures in
natural populations of the Trinidadian guppy, Poecilia reticulata.
Behav Ecol Sociobiol 55:461–468
Plaistow S, Siva-Jothy MT (1996) Energetic constraints and male
mate-securing tactics in the damselfly Calopteryx splendens
xanthostoma (Charpentier). Proc R Soc Lond B 263:1233–1239
Price AC, Rodd FH (2006) The effect of social environment on male–
male competition in guppies (Poecilia reticulata). Ethology
112:22–32
Price TD, Qvarnstrom A, Irwin DE (2003) The role of phenotypic
plasticity in driving genetic evolution. Proc R Soc Lond B
270:1433–1440
Quinn TP, Adkison MD, Ward MB (1996) Behavioral tactics of male
sockeye salmon (Oncorhynchus nerka) under varying operational
sex ratios. Ethology 102:304–322
Behav Ecol Sociobiol (2007) 61:689–701
Reyer H-U, Fischer W, Steck P, Nabulon T, Kessler P (1998) Sexspecific nest defense in house sparrows (Passer domesticus)
varies with badge size of males. Behav Ecol Sociobiol 42:93–99
Reynolds JD, Gross MR, Coombs J (1993) Environmental conditions
and male morphology determine alternative mating behavior in
Trinidadian guppies. Anim Behav 45:145–152
Reznick D, Butler MJ IV, Rodd H (2001) Life-history evolution in
guppies. VII. The comparative ecology of high- and lowpredation environments. Am Nat 157:126–140
Rodd FH, Reznick DN (1991) Life-history evolution in guppies.3: the
impact of prawn predation on guppy life histories. Oikos 62:13–19
Rodd FH, Sokolowski MB (1995) Complex origins of variation in the
sexual behavior of male Trinidadian guppies, Poecilia reticulata:
interactions between social environment, heredity, body size and
age. Anim Behav 49:1139–1159
Rodd FH, Reznick DN, Sokolowski MB (1997) Phenotypic plasticity
in the life history traits of guppies: responses to social
environment. Ecology 78:419–433
Rotenberry JT, Wiens JA (1998) Foraging patch selection by
shrubsteppe sparrows. Ecology 79:1160–1173
Shuster SM, Wade MJ (2003) Mating systems and strategies. Princeton
University Press, Princeton
701
Tabachnick BG, Fidell LS (2001) Using multivariate statistics, 4th
edn. Allyn and Bacon, Needham Heights, MA
Thompson DB (1999) Different spatial scales of natural selection and
gene flow: the evolution of behavioral geographic variation and
phenotypic plasticity. In: Foster SA, Endler JA (eds) Geographic
variation in behavior. Oxford University Press, New York, pp 52–68
Templeton CN, Shriner WM (2004) Multiple selection pressures
influence Trinidadian guppy (Poecilia reticulata) antipredator
behavior. Behav Ecol 4:673–678
Verhoeven KJF, Simonsen KL, McIntyre LM (2005) Implementing
false discovery rate control: increasing your power. Oikos
108:643–647
Weitere M, Tautz D, Neumann D, Steinfartz S (2004) Adaptive
divergence vs. environmental plasticity: tracing local genetic
adaptation of metamorphosis traits in salamanders. Mol Ecol
13:1665–1677
West-Eberhard MJ (2003) Developmental Plasticity and Evolution.
Oxford University Press, Oxford
Winge O (1937) Succession of broods in Lebistes. Nature 140:467
Zuk M, Rotenberry JT, Simmons LW (1998) Calling songs of field
crickets (Teleogryllus oceanicus) with and without phonotactic
parasitoid infection. Evolution 52:166–171